From 2f335a93bf70c067f3029fbc5445607df7552e0c Mon Sep 17 00:00:00 2001
From: nikosbosse <nikosbosse@gmail.com>
Date: Tue, 19 Dec 2023 17:53:34 +0100
Subject: [PATCH 1/5] Rename forecast classes to `forecast_*`

---
 NAMESPACE                             | 16 ++++++++--------
 R/score.R                             | 14 ++++++++------
 R/validate.R                          | 18 +++++++++---------
 man/as_forecast.Rd                    |  4 ++--
 man/score.Rd                          | 16 ++++++++--------
 man/validate_forecast.Rd              |  4 ++--
 tests/testthat/test-check_forecasts.R |  2 +-
 tests/testthat/test-metrics-point.R   |  4 ++--
 tests/testthat/test-score.R           |  6 +++---
 9 files changed, 43 insertions(+), 41 deletions(-)

diff --git a/NAMESPACE b/NAMESPACE
index c41ef7009..eb38a7e15 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -5,14 +5,14 @@ S3method(print,scoringutils_check)
 S3method(quantile_to_interval,data.frame)
 S3method(quantile_to_interval,numeric)
 S3method(score,default)
-S3method(score,scoringutils_binary)
-S3method(score,scoringutils_point)
-S3method(score,scoringutils_quantile)
-S3method(score,scoringutils_sample)
-S3method(validate_forecast,scoringutils_binary)
-S3method(validate_forecast,scoringutils_point)
-S3method(validate_forecast,scoringutils_quantile)
-S3method(validate_forecast,scoringutils_sample)
+S3method(score,forecast_binary)
+S3method(score,forecast_point)
+S3method(score,forecast_quantile)
+S3method(score,forecast_sample)
+S3method(validate_forecast,forecast_binary)
+S3method(validate_forecast,forecast_point)
+S3method(validate_forecast,forecast_quantile)
+S3method(validate_forecast,forecast_sample)
 export(abs_error)
 export(add_coverage)
 export(add_pairwise_comparison)
diff --git a/R/score.R b/R/score.R
index 30d717b1a..d585593ee 100644
--- a/R/score.R
+++ b/R/score.R
@@ -70,13 +70,13 @@ score <- function(data, ...) {
 score.default <- function(data, ...) {
   assert(check_data_columns(data))
   forecast_type <- get_forecast_type(data)
-  data <- new_scoringutils(data, paste0("scoringutils_", forecast_type))
+  data <- new_scoringutils(data, paste0("forecast_", forecast_type))
   score(data, ...)
 }
 
 #' @rdname score
 #' @export
-score.scoringutils_binary <- function(data, metrics = metrics_binary, ...) {
+score.forecast_binary <- function(data, metrics = metrics_binary, ...) {
   data <- validate_forecast(data)
   data <- remove_na_observed_predicted(data)
   metrics <- validate_metrics(metrics)
@@ -96,7 +96,7 @@ score.scoringutils_binary <- function(data, metrics = metrics_binary, ...) {
 #' @importFrom Metrics se ae ape
 #' @rdname score
 #' @export
-score.scoringutils_point <- function(data, metrics = metrics_point, ...) {
+score.forecast_point <- function(data, metrics = metrics_point, ...) {
   data <- validate_forecast(data)
   data <- remove_na_observed_predicted(data)
   metrics <- validate_metrics(metrics)
@@ -113,7 +113,7 @@ score.scoringutils_point <- function(data, metrics = metrics_point, ...) {
 
 #' @rdname score
 #' @export
-score.scoringutils_sample <- function(data, metrics = metrics_sample, ...) {
+score.forecast_sample <- function(data, metrics = metrics_sample, ...) {
   data <- validate_forecast(data)
   data <- remove_na_observed_predicted(data)
   forecast_unit <- attr(data, "forecast_unit")
@@ -150,7 +150,7 @@ score.scoringutils_sample <- function(data, metrics = metrics_sample, ...) {
 #' @importFrom data.table `:=` as.data.table rbindlist %like%
 #' @rdname score
 #' @export
-score.scoringutils_quantile <- function(data, metrics = metrics_quantile, ...) {
+score.forecast_quantile <- function(data, metrics = metrics_quantile, ...) {
   data <- validate_forecast(data)
   data <- remove_na_observed_predicted(data)
   forecast_unit <- attr(data, "forecast_unit")
@@ -176,7 +176,9 @@ score.scoringutils_quantile <- function(data, metrics = metrics_quantile, ...) {
     observed <- data$observed
     predicted <- do.call(rbind, data$predicted)
     quantile <- unlist(unique(data$quantile))
-    data[, c("observed", "predicted", "quantile", "scoringutils_quantile") := NULL]
+    data[, c(
+      "observed", "predicted", "quantile", "scoringutils_quantile"
+    ) := NULL]
 
     data <- apply_metrics(
       data, metrics,
diff --git a/R/validate.R b/R/validate.R
index 65da31eba..20d3c9197 100644
--- a/R/validate.R
+++ b/R/validate.R
@@ -12,8 +12,8 @@
 #' @inheritParams score
 #' @inheritSection forecast_types Forecast types and input format
 #' @return Depending on the forecast type, an object of class
-#' `scoringutils_binary`, `scoringutils_point`, `scoringutils_sample` or
-#' `scoringutils_quantile`.
+#' `forecast_binary`, `forecast_point`, `forecast_sample` or
+#' `forecast_quantile`.
 #' @export
 #' @keywords check-forecasts
 #' @examples
@@ -32,7 +32,7 @@ as_forecast.default <- function(data, ...) {
   forecast_type <- get_forecast_type(data)
 
   # construct class
-  data <- new_scoringutils(data, paste0("scoringutils_", forecast_type))
+  data <- new_scoringutils(data, paste0("forecast_", forecast_type))
 
   # validate class
   validate_forecast(data)
@@ -48,8 +48,8 @@ as_forecast.default <- function(data, ...) {
 #' @inheritParams score
 #' @inheritSection forecast_types Forecast types and input format
 #' @return Depending on the forecast type, an object of class
-#' `scoringutils_binary`, `scoringutils_point`, `scoringutils_sample` or
-#' `scoringutils_quantile`.
+#' `forecast_binary`, `forecast_point`, `forecast_sample` or
+#' `forecast_quantile`.
 #' @importFrom data.table ':=' is.data.table
 #' @importFrom checkmate assert_data_frame
 #' @export
@@ -65,7 +65,7 @@ validate_forecast <- function(data, ...) {
 #' @rdname validate
 #' @export
 #' @keywords check-forecasts
-validate_forecast.scoringutils_binary <- function(data, ...) {
+validate_forecast.forecast_binary <- function(data, ...) {
   data <- validate_general(data)
 
   columns_correct <- test_columns_not_present(data, c("sample_id", "quantile"))
@@ -86,7 +86,7 @@ validate_forecast.scoringutils_binary <- function(data, ...) {
 #' @rdname validate
 #' @export
 #' @keywords check-forecasts
-validate_forecast.scoringutils_point <- function(data, ...) {
+validate_forecast.forecast_point <- function(data, ...) {
   data <- validate_general(data)
 
   input_check <- check_input_point(data$observed, data$predicted)
@@ -100,7 +100,7 @@ validate_forecast.scoringutils_point <- function(data, ...) {
 
 #' @rdname validate
 #' @export
-validate_forecast.scoringutils_quantile <- function(data, ...) {
+validate_forecast.forecast_quantile <- function(data, ...) {
   data <- validate_general(data)
   assert_numeric(data$quantile, lower = 0, upper = 1)
   return(data[])
@@ -109,7 +109,7 @@ validate_forecast.scoringutils_quantile <- function(data, ...) {
 #' @rdname validate
 #' @export
 #' @keywords check-forecasts
-validate_forecast.scoringutils_sample <- function(data, ...) {
+validate_forecast.forecast_sample <- function(data, ...) {
   data <- validate_general(data)
   return(data[])
 }
diff --git a/man/as_forecast.Rd b/man/as_forecast.Rd
index 11e6091ce..b9f3a1b47 100644
--- a/man/as_forecast.Rd
+++ b/man/as_forecast.Rd
@@ -16,8 +16,8 @@ as_forecast(data, ...)
 }
 \value{
 Depending on the forecast type, an object of class
-\code{scoringutils_binary}, \code{scoringutils_point}, \code{scoringutils_sample} or
-\code{scoringutils_quantile}.
+\code{forecast_binary}, \code{forecast_point}, \code{forecast_sample} or
+\code{forecast_quantile}.
 }
 \description{
 Convert a data.frame or similar of forecasts into an object of
diff --git a/man/score.Rd b/man/score.Rd
index 92b484bed..d317ed64a 100644
--- a/man/score.Rd
+++ b/man/score.Rd
@@ -3,23 +3,23 @@
 \name{score}
 \alias{score}
 \alias{score.default}
-\alias{score.scoringutils_binary}
-\alias{score.scoringutils_point}
-\alias{score.scoringutils_sample}
-\alias{score.scoringutils_quantile}
+\alias{score.forecast_binary}
+\alias{score.forecast_point}
+\alias{score.forecast_sample}
+\alias{score.forecast_quantile}
 \title{Evaluate forecasts in a data.frame format}
 \usage{
 score(data, ...)
 
 \method{score}{default}(data, ...)
 
-\method{score}{scoringutils_binary}(data, metrics = metrics_binary, ...)
+\method{score}{forecast_binary}(data, metrics = metrics_binary, ...)
 
-\method{score}{scoringutils_point}(data, metrics = metrics_point, ...)
+\method{score}{forecast_point}(data, metrics = metrics_point, ...)
 
-\method{score}{scoringutils_sample}(data, metrics = metrics_sample, ...)
+\method{score}{forecast_sample}(data, metrics = metrics_sample, ...)
 
-\method{score}{scoringutils_quantile}(data, metrics = metrics_quantile, ...)
+\method{score}{forecast_quantile}(data, metrics = metrics_quantile, ...)
 }
 \arguments{
 \item{data}{A data.frame or data.table with predicted and observed values.}
diff --git a/man/validate_forecast.Rd b/man/validate_forecast.Rd
index 5d4bbe2d7..542e84dd5 100644
--- a/man/validate_forecast.Rd
+++ b/man/validate_forecast.Rd
@@ -13,8 +13,8 @@ validate_forecast(data, ...)
 }
 \value{
 Depending on the forecast type, an object of class
-\code{scoringutils_binary}, \code{scoringutils_point}, \code{scoringutils_sample} or
-\code{scoringutils_quantile}.
+\code{forecast_binary}, \code{forecast_point}, \code{forecast_sample} or
+\code{forecast_quantile}.
 }
 \description{
 Methods for the different classes run \code{\link[=validate_general]{validate_general()}}, which performs
diff --git a/tests/testthat/test-check_forecasts.R b/tests/testthat/test-check_forecasts.R
index 1441c3f7f..004031d71 100644
--- a/tests/testthat/test-check_forecasts.R
+++ b/tests/testthat/test-check_forecasts.R
@@ -1,6 +1,6 @@
 test_that("as_forecast() function works", {
   check <- suppressMessages(as_forecast(example_quantile))
-  expect_s3_class(check, "scoringutils_quantile")
+  expect_s3_class(check, "forecast_quantile")
 })
 
 test_that("as_forecast() function has an error for empty data.frame", {
diff --git a/tests/testthat/test-metrics-point.R b/tests/testthat/test-metrics-point.R
index 2f64226df..54aeb623c 100644
--- a/tests/testthat/test-metrics-point.R
+++ b/tests/testthat/test-metrics-point.R
@@ -179,9 +179,9 @@ test_that("abs error is correct, point and median forecasts same", {
     observations = truth_scoringutils
   )
 
-  data_scoringutils_point <- data_scoringutils[type == "point"][, quantile := NULL]
+  data_forecast_point <- data_scoringutils[type == "point"][, quantile := NULL]
 
-  eval <- score(data = data_scoringutils_point)
+  eval <- score(data = data_forecast_point)
   eval <- summarise_scores(eval,
     by = c(
       "location", "target_end_date",
diff --git a/tests/testthat/test-score.R b/tests/testthat/test-score.R
index a67575984..29132de21 100644
--- a/tests/testthat/test-score.R
+++ b/tests/testthat/test-score.R
@@ -38,7 +38,7 @@ test_that("function produces output for a binary case", {
   expect_true("brier_score" %in% names(eval))
 })
 
-test_that("score.scoringutils_binary() errors with only NA values", {
+test_that("score.forecast_binary() errors with only NA values", {
   only_nas <- copy(example_binary)[, predicted := NA_real_]
   expect_error(
     score(only_nas),
@@ -156,7 +156,7 @@ test_that("Changing metrics names works", {
 })
 
 
-test_that("score.scoringutils_point() errors with only NA values", {
+test_that("score.forecast_point() errors with only NA values", {
   only_nas <- copy(example_point)[, predicted := NA_real_]
   expect_error(
     score(only_nas),
@@ -239,7 +239,7 @@ test_that("WIS is the same with other metrics omitted or included", {
 })
 
 
-test_that("score.scoringutils_quantile() errors with only NA values", {
+test_that("score.forecast_quantile() errors with only NA values", {
   only_nas <- copy(example_quantile)[, predicted := NA_real_]
   expect_error(
     score(only_nas),

From d58209e0e61f96ed8c8f977e5dcdd257fbee3e5a Mon Sep 17 00:00:00 2001
From: nikosbosse <nikosbosse@gmail.com>
Date: Tue, 19 Dec 2023 17:55:50 +0100
Subject: [PATCH 2/5] rename `new_scoringutils()` to `new_forecast()`

---
 NAMESPACE                                    |  2 +-
 R/score.R                                    |  2 +-
 R/validate.R                                 | 12 ++++++------
 man/as_forecast.Rd                           |  2 +-
 man/{new_scoringutils.Rd => new_forecast.Rd} |  6 +++---
 5 files changed, 12 insertions(+), 12 deletions(-)
 rename man/{new_scoringutils.Rd => new_forecast.Rd} (88%)

diff --git a/NAMESPACE b/NAMESPACE
index eb38a7e15..35c1eb7c0 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -41,7 +41,7 @@ export(mad_sample)
 export(make_NA)
 export(make_na)
 export(merge_pred_and_obs)
-export(new_scoringutils)
+export(new_forecast)
 export(overprediction)
 export(pairwise_comparison)
 export(pit)
diff --git a/R/score.R b/R/score.R
index d585593ee..524d2cbfc 100644
--- a/R/score.R
+++ b/R/score.R
@@ -70,7 +70,7 @@ score <- function(data, ...) {
 score.default <- function(data, ...) {
   assert(check_data_columns(data))
   forecast_type <- get_forecast_type(data)
-  data <- new_scoringutils(data, paste0("forecast_", forecast_type))
+  data <- new_scoringutils(data, paste0("forecast_ ", forecast_type))
   score(data, ...)
 }
 
diff --git a/R/validate.R b/R/validate.R
index 20d3c9197..814e396fc 100644
--- a/R/validate.R
+++ b/R/validate.R
@@ -6,7 +6,7 @@
 #' quantile-based) from the input data (using the function
 #' [get_forecast_type()]. It then constructs an object of the
 #' appropriate class (`forecast_binary`, `forecast_point`, `forecast_sample`, or
-#' `forecast_quantile`, using the function [new_scoringutils()]).
+#' `forecast_quantile`, using the function [new_forecast()]).
 #' Lastly, it calls [as_forecast()] on the object to make sure it conforms with
 #' the required input formats.
 #' @inheritParams score
@@ -32,7 +32,7 @@ as_forecast.default <- function(data, ...) {
   forecast_type <- get_forecast_type(data)
 
   # construct class
-  data <- new_scoringutils(data, paste0("forecast_", forecast_type))
+  data <- new_forecast(data, paste0("forecast_", forecast_type))
 
   # validate class
   validate_forecast(data)
@@ -62,7 +62,6 @@ validate_forecast <- function(data, ...) {
 }
 
 
-#' @rdname validate
 #' @export
 #' @keywords check-forecasts
 validate_forecast.forecast_binary <- function(data, ...) {
@@ -83,7 +82,7 @@ validate_forecast.forecast_binary <- function(data, ...) {
   return(data[])
 }
 
-#' @rdname validate
+
 #' @export
 #' @keywords check-forecasts
 validate_forecast.forecast_point <- function(data, ...) {
@@ -106,7 +105,7 @@ validate_forecast.forecast_quantile <- function(data, ...) {
   return(data[])
 }
 
-#' @rdname validate
+
 #' @export
 #' @keywords check-forecasts
 validate_forecast.forecast_sample <- function(data, ...) {
@@ -114,6 +113,7 @@ validate_forecast.forecast_sample <- function(data, ...) {
   return(data[])
 }
 
+
 #' @title Apply scoringutls input checks that are the same across forecast types
 #'
 #' @description
@@ -181,7 +181,7 @@ validate_general <- function(data) {
 #' @return An object of the class indicated by `classname`
 #' @export
 #' @keywords internal
-new_scoringutils <- function(data, classname) {
+new_forecast <- function(data, classname) {
   data <- as.data.table(data)
   data <- assure_model_column(data)
   class(data) <- c(classname, class(data))
diff --git a/man/as_forecast.Rd b/man/as_forecast.Rd
index b9f3a1b47..3704364a8 100644
--- a/man/as_forecast.Rd
+++ b/man/as_forecast.Rd
@@ -27,7 +27,7 @@ class \verb{forecast_*} and validate it.
 quantile-based) from the input data (using the function
 \code{\link[=get_forecast_type]{get_forecast_type()}}. It then constructs an object of the
 appropriate class (\code{forecast_binary}, \code{forecast_point}, \code{forecast_sample}, or
-\code{forecast_quantile}, using the function \code{\link[=new_scoringutils]{new_scoringutils()}}).
+\code{forecast_quantile}, using the function \code{\link[=new_forecast]{new_forecast()}}).
 Lastly, it calls \code{\link[=as_forecast]{as_forecast()}} on the object to make sure it conforms with
 the required input formats.
 }
diff --git a/man/new_scoringutils.Rd b/man/new_forecast.Rd
similarity index 88%
rename from man/new_scoringutils.Rd
rename to man/new_forecast.Rd
index e19feb989..772c5e1b4 100644
--- a/man/new_scoringutils.Rd
+++ b/man/new_forecast.Rd
@@ -1,10 +1,10 @@
 % Generated by roxygen2: do not edit by hand
 % Please edit documentation in R/validate.R
-\name{new_scoringutils}
-\alias{new_scoringutils}
+\name{new_forecast}
+\alias{new_forecast}
 \title{Class constructor for scoringutils objects}
 \usage{
-new_scoringutils(data, classname)
+new_forecast(data, classname)
 }
 \arguments{
 \item{data}{A data.frame or data.table with predicted and observed values.}

From ded04f5815b8df7eeb85ee43b494b8c541b04c7b Mon Sep 17 00:00:00 2001
From: nikosbosse <nikosbosse@gmail.com>
Date: Tue, 19 Dec 2023 17:57:01 +0100
Subject: [PATCH 3/5] change new_scoringutils to new_forecast

---
 R/score.R | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/R/score.R b/R/score.R
index 524d2cbfc..ef93a0d2c 100644
--- a/R/score.R
+++ b/R/score.R
@@ -70,7 +70,7 @@ score <- function(data, ...) {
 score.default <- function(data, ...) {
   assert(check_data_columns(data))
   forecast_type <- get_forecast_type(data)
-  data <- new_scoringutils(data, paste0("forecast_ ", forecast_type))
+  data <- new_forecast(data, paste0("forecast_ ", forecast_type))
   score(data, ...)
 }
 

From 5a1712365ec3bfd3f916e90ad56eafb21a82c984 Mon Sep 17 00:00:00 2001
From: nikosbosse <nikosbosse@gmail.com>
Date: Tue, 19 Dec 2023 18:14:49 +0100
Subject: [PATCH 4/5] Rename instances of "scoringutils_" to "forecast_" when
 constructing new objects

---
 R/get_-functions.R                 | 2 +-
 R/score.R                          | 2 +-
 R/validate.R                       | 3 ++-
 man/get_scoringutils_attributes.Rd | 2 +-
 4 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/R/get_-functions.R b/R/get_-functions.R
index e889911b4..f68434651 100644
--- a/R/get_-functions.R
+++ b/R/get_-functions.R
@@ -250,7 +250,7 @@ get_duplicate_forecasts <- function(data, forecast_unit = NULL) {
 
 #' @title Get a list of all attributes of a scoringutils object
 #'
-#' @param object A object of class `scoringutils_`
+#' @param object A object of class `forecast_`
 #'
 #' @return A named list with the attributes of that object.
 #' @keywords internal
diff --git a/R/score.R b/R/score.R
index ef93a0d2c..6c1298d90 100644
--- a/R/score.R
+++ b/R/score.R
@@ -70,7 +70,7 @@ score <- function(data, ...) {
 score.default <- function(data, ...) {
   assert(check_data_columns(data))
   forecast_type <- get_forecast_type(data)
-  data <- new_forecast(data, paste0("forecast_ ", forecast_type))
+  data <- new_forecast(data, paste0("forecast_", forecast_type))
   score(data, ...)
 }
 
diff --git a/R/validate.R b/R/validate.R
index 814e396fc..4a1be5f1f 100644
--- a/R/validate.R
+++ b/R/validate.R
@@ -97,8 +97,9 @@ validate_forecast.forecast_point <- function(data, ...) {
   return(data[])
 }
 
-#' @rdname validate
+
 #' @export
+#' @keywords check-forecasts
 validate_forecast.forecast_quantile <- function(data, ...) {
   data <- validate_general(data)
   assert_numeric(data$quantile, lower = 0, upper = 1)
diff --git a/man/get_scoringutils_attributes.Rd b/man/get_scoringutils_attributes.Rd
index 82ce6dcbe..5f2770a6a 100644
--- a/man/get_scoringutils_attributes.Rd
+++ b/man/get_scoringutils_attributes.Rd
@@ -7,7 +7,7 @@
 get_scoringutils_attributes(object)
 }
 \arguments{
-\item{object}{A object of class \code{scoringutils_}}
+\item{object}{A object of class \code{forecast_}}
 }
 \value{
 A named list with the attributes of that object.

From 93094412393c1f5ff4a7e6ab05c65142d68345f0 Mon Sep 17 00:00:00 2001
From: nikosbosse <nikosbosse@gmail.com>
Date: Tue, 19 Dec 2023 18:28:42 +0100
Subject: [PATCH 5/5] Update News.md

---
 NEWS.md | 1 +
 1 file changed, 1 insertion(+)

diff --git a/NEWS.md b/NEWS.md
index b7fd1d529..e77c172c7 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -6,6 +6,7 @@ The update introduces breaking changes. If you want to keep using the older vers
 
 ## Package updates
 - In `score()`, required columns "true_value" and "prediction" were renamed and replaced by required columns "observed" and "predicted". Scoring functions now also use the function arguments "observed" and "predicted" everywhere consistently. 
+- The overall scoring workflow was updated. `score()` is now a generic function that dispatches the correct method based on the forecast type. forecast types currently supported are "binary", "point", "sample" and "quantile" with corresponding classes "forecast_binary", "forecast_point", "forecast_sample" and "forecast_quantile". An object of class `forecast_*` can be created using the function `as_forecast()`, which also replaces the previous function `check_forecasts()` (see more information below). 
 - Scoring functions received a consistent interface and input checks:
   - metrics for binary forecasts:
     - `observed`: factor with exactly 2 levels